Hydrogen effects in hexane reactions over Al2O3 supported Pt, Ir and Pt-Ir catalysts

Article abstract of DOI:10.1039/a902976a

Hexane was tested under similar conditions on Al₂O₃ supported Pt, Ir and Pt-Ir catalysts. A conventional impregnation on commercial alumina and a sol-gel procedure involving the introduction of the Pt or Ir components at the commencement of gelification were applied for the preparation of monometallic catalysts. Pt-It samples were prepared by co-impregnation or using a bimetallic complex precursor, [Ir(NH₃)₅Cl][PtCl₄], produced in the pores of the support. Increasing H₂ pressure at both 603 and 663 K enhanced the overall activity and the very pronounced hydrogenolysis on Ir. Non- degradative products, including isomers, methylcyclopentane, benzene and hexenes, showed, as a rule, maximum rates as a function of p(H₂) on Ir also, in agreement with earlier and present reports concerning Pt. Both Pt and Ir catalysts on a sol-gel alumina were superior in non-degradative reactions to their conventional counterparts. The hydrogenolysis activity of both bimetallic catalysts was below the average for Pt and Ir, indicating Pt-Ir interactions. These were more pronounced with Pt-Ir prepared from a complex precursor, showing a higher activity in non-degradative reactions, more manifest at 663 K.

Full text of DOI:10.1039/a902976a

View Article Online / Journal Homepage / Table of Contents for this issue
Hydrogen e†ects in hexane reactions over Al O supported Pt, Ir and
2 3
Pt–Ir catalysts
Aurelie Charron,ab Charles Kappenstein,*a Maurice Guerina and Zoltan Paalb
a L ACCO, UMR-CNRS 6503, Catalyse en Chimie Organique, Faculte des Sciences, 86022
Poitiers Cedex, France
b Institute of Isotope and Surface Chemistry, CRC, Hungarian Academy of Sciences, P.O. Box 77,
H-1525 Budapest, Hungary
Received 14th April 1999, Accepted 25th June 1999
Hexane was tested under similar conditions on Al O supported Pt, Ir and PtÈIr catalysts. A conventional
2
3
impregnation on commercial alumina and a solÈgel procedure involving the introduction of the Pt or Ir
components at the commencement of geliÐcation were applied for the preparation of monometallic catalysts.
PtÈIr samples were prepared by co-impregnation or using a bimetallic complex precursor,
[Ir(NH ) Cl][PtCl ], produced in the pores of the support. Increasing H pressure at both 603 and 663 K
3 5
4
2
enhanced the overall activity and the very pronounced hydrogenolysis on Ir. Non-degradative products,
including isomers, methylcyclopentane, benzene and hexenes, showed, as a rule, maximum rates as a function
of p(H ) on Ir also, in agreement with earlier and present reports concerning Pt. Both Pt and Ir catalysts on a
2
solÈgel alumina were superior in non-degradative reactions to their conventional counterparts. The
hydrogenolysis activity of both bimetallic catalysts was below the average for Pt and Ir, indicating PtÈIr
interactions. These were more pronounced with PtÈIr prepared from a complex precursor, showing a higher
activity in non-degradative reactions, more manifest at 663 K.
Skeletal reactions of alkanes are important processes for
naphtha reforming.1,2 They include dehydrocyclization of
alkanes into aromatics, along with their skeletal isomerization
successfully to do the same with PtÈSn catalysts prepared
from di†erent precursors.15h17
One of the novelties of the above method of preparation of
bimetallic catalysts was the use of a new, complex precursor
containing the two metallic components in a ratio of 1 : 1. It
was expected, and indeed found, that the catalysts prepared in
this way would give a more even distribution of the individual
components. The use of such a complex precursor for PtÈIr
catalysts was Ðrst reported by Michelot et al.18 The same pro-
cedure was applied to prepare Al O supported PtÈIr cata-
and C cyclization. At the temperature of commercial reform-
5
ing, these latter processes occur by cooperation of metallic
and acidic sites, but there is no doubt that metals themselves
are able to catalyse these reactions.1
Platinum is used most widely for these reactions and its
bimetallic combinations, such as PtÈSn, PtÈRe and PtÈIr, also
have some importance.3 Some of these additives have no cata-
lytic properties on their own (such as Sn or, to some extent,
Re), whereas Ir can be a very active catalyst of hydrocarbon
transformations.4 As opposed to Pt, iridium exhibited a much
higher hydrogenolysis activity as a single crystal,5 a foil6 or a
disperse metal.7h9 The properties of PtÈIr on Al O supports
2
3
lysts. In this work, the catalytic results are compared with
those obtained over a traditionally prepared PtÈIr/Al O
2
3
catalyst and monometallic samples of similar origin. Mono-
metallic Pt/Al O and Ir/Al O samples prepared by the new
2
3
2 3
solÈgel technique were also included into the catalytic testing
2
3
have been extensively studied by Rasser et al.10 The high
hydrogenolysis activity of Ir manifested itself in hindered
coking, probably by splitting of hydrocarbonaceous coke pre-
cursors. They also found a surface enrichment of Pt up to high
Ir contents, in agreement with other reports and calculations
based on solid-state physics.3,6,11
procedure.
Earlier studies on alkane hydrogenolysis over Ir (and also
Pt) catalysts reported a very pronounced hampering e†ect of
increasing hydrogen pressure.19 Curves with maxima were
typical, in turn, for yields of skeletal reactions on Pt cata-
lysts.1,20,21 This paper reports experiments to compare the Pt,
Ir and PtÈIr catalysts under identical conditions.
The isomerization propensity of Ir necessarily plays a minor
role compared with hydrogenolysis, in contrast to Pt.12 Ir was
found to be among the four metals (the others being Pt, Pd
and Rh) that are able to produce skeletal isomers from 3-
Experimental
Catalyst preparation
methylpentane,13 probably involving the ““C -cyclic routeÏÏ of
5
isomerization. The Ðrst reports of isomerization on Ir claimed
All the catalysts had a metal loading of 1% w/w and were
prepared with Al O GFSC supplied by Rhone-Poulenc (RP,
that the ““C -cyclic intermediateÏÏ could only be formed
5
2 3
between primary C atoms.14 However, other studies demon-
France) as the support, except for MS3 and MS4. The support
(surface area 210 m2 g~1; porous volume 0.64 cm3 g~1) was
ground and sieved (0.1È0.25 mm), then washed with distilled
water, dried at 373 K, calcined in an air Ñow at 800 K (4 h)
strated that Ir was also able to produce both methyl-
cyclopentane (MCP) and skeletal isomers from hexane12 at
lower temperature and toluene from heptane5 at higher tem-
peratures. Hence, the reactions of hexane seemed to be an
appropriate test for the characterisation of the catalytic
properties of Ir containing catalysts. This reaction was used
and Ðnally reduced in a stream of H at 800 K (4 h) before
2
use. The impregnation procedure will be illustrated on the
example of PtÈIr/Al O samples (Pt : Ir \ 1 : 1, i.e., 0.5% w/w
2
3
Phys. Chem. Chem. Phys., 1999, 1, 3817È3822
3817

View Article Online
Ir and 0.5% w/w Pt). The conventional preparation (MS1)
involved co-impregnation of 4 g of Al O with a mixture of
catalysts took an intermediate position. The activity data for
Pt/Al O (RP) (MS7) were taken from ref. 16. The various
2
3
2 3
ca. 3 ] 3 mL of aqueous solution containing 35 mmol L~1
preparation methods a†ected the activities in di†erent ways.
The solÈgel preparation produced a more active Pt but a less
active the Ir catalyst. However, while the activity of the
Ir/Al O (RP) catalyst levelled o† above about p(H ) \ 32
each of H PtCl and H IrCl . The support was Ðrst covered
2
6
2
6
with water before adding the above solution of metal salts.
The slurry was slowly dried in a rotational evaporator for 3 h
at 333 K at water pump pressure and then in an oven at 393
K overnight. The dispersion was 44%, as determined by con-
ventional hydrogenÈoxygen titration.
2
3
2
kPa, that of the solÈgel preparation increased continuously
and one could extrapolate that it might be most active at even
higher hydrogen pressures.
For sample MS2, the formation of the bimetallic precursor
[Ir(NH ) Cl][PtCl ] in the pores of the alumina needs two
Comparing the speciÐc activities (per unit mass of metal) of
the two monometallic catalysts in the formation of individual
products (Table 2), it is obvious that the activity increase of Ir
catalysts, as shown in Table 1, was exclusively due to the
larger amounts of fragments. The di†erence could reach a
factor of 10, which is certainly far below the di†erences of 5È6
orders of magnitude reported earlier.4,19 The activity of Pt
catalysts towards benzene was higher by a factor of D2; at
3 5
4
steps. First, the [Ir(NH ) Cl]2` cationic complex was intro-
3 5
duced by two successive incipient wetness impregnations using
an aqueous solution (typically 2.6 cm3 each) containing the
calculated amount of iridium salt, [Ir(NH ) Cl]Cl , for 4 g of
3 5
2
alumina. After stirring for 15 min with the rotational evapo-
rator, the sample was slightly dried at 333 K in order to pre-
serve the integrity of the iridium species. The introduction of
the complex anion from (NH ) PtCl solution followed, using
the same time the rates of isomerization and C cyclization
5
were larger by about 5È10-fold on Pt. The rates showed
4 2
4
the same two-step impregnation procedure. The PtÈIr precur-
sor develops a yellowÈorange colour on the support prior to
reduction at 773 K. The dispersion after reduction was 26%.
MS3 was Pt/Al O and MS4 was Ir/Al O , prepared by
maxima as a function of the hydrogen pressure for both cata-
lyst metals and for both preparations of Ir. Pt on solÈgel
alumina, in turn, exhibited no maximum rate. The solÈgel
method of preparation gave less active catalysts for C satu-
2
3
2
3
6
the solÈgel method, introducing the Pt or Ir salts at the com-
mencement of the solÈgel process. The solÈgel preparation
involved the hydrolysis of aluminium tri(sec-butoxide),
Al(OC H ) , in the presence of dilute HCl, following the pro-
rated products (isomers, MCP), the di†erence being more pro-
nounced with Pt, this sample also giving more benzene.
The intrinsic activities of the bimetallic PtÈIr catalysts were
fairly close to each other. The turnover frequency (TOF) value
for MS2 (new precursor) was higher by a factor of only about
1.5 times that of the conventional impregnation (Table 3). This
di†erence was maintained at 633 K.
Fragments, benzene, skeletal isomers, methylcyclopentane
(having undergone skeletal rearrangement) and hexenes were
among the products on all catalysts. Their abundances were
very di†erent on di†erent metals. The hydrogen pressure
dependence of these products on both PtÈIr catalysts is
included in Table 3. Increasing the hydrogen pressure from 8
to 64 kPa increased the fragmentation activity by a factor of
about 20. Although the other products were formed in minor
amounts, maxima appeared in their yields as a function of
4
9 3
cedure described by Lopez et al.22 for silicoaluminates.
Results for Pt/Al O (RP) (sample MS7) were taken from ref.
16 and compared with an Ir/Al O catalyst of similar prep-
aration (MS8). The dispersions of MS7 and MS8 were 32 and
54%, respectively. Activity results obtained on a well charac-
2
3
2
3
terised 6% Pt/SiO catalyst23 (EUROPT-1) were also
2
included for comparison.
Catalytic runs
A closed-loop circulation reactor described earlier15,16 was
used for catalytic tests. A pre-mixed blend of 10 kPa hexane
plus H at various pressures (between 8 and 64 kPa) was pre-
p(H ) for all other skeletal products. Their shift towards
2
2
pared and introduced to the reactor vessel (around 200 mL)
higher p(H ) at 663 K corresponded to earlier experience with
2
containing 50 mg of catalyst. The reactions were studied at
two temperatures, 603 and 663 K, and the products were
analysed by gas chromatography.15,16 The selectivities were
Pt. The amount of hexenes always decreased as a function of
the hydrogen pressure.
Of the two PtÈIr samples, that prepared from a bimetallic
precursor showed a higher intrinsic activity. This catalyst also
produced more benzene and also alkenes under identical con-
ditions. This propensity is analogous to that observed with the
PtÈSn ex-bimetallic complex where it was attributed to a Ðner
distribution of Pt and Sn atoms. The ““small ensemblesÏÏ
created in this way were regarded as responsible for non-
degradative alkane reactions.24 If these ensembles consist pre-
dominantly of single metal atoms, dehydrogenation to
hexenes25 and benzene formation should prevail, the latter
involving a further stepwise loss of H atoms.1,26 At 663 K
calculated for C units reacted, rather than moles of products
6
formed.
Results
Catalytic activity
The speciÐc activity of all catalysts increased monotonically
with increase in hydrogen pressures (Table 1). Of all the cata-
lysts, the values for monometallic Ir samples were the highest,
those for monometallic Pt were the lowest and the bimetallic
Table 1 Overall activity of Pt, Ir and PtÈIr catalysts as a function of the hydrogen pressure
SpeciÐc activity/10~4 mol (mg metal)~1 h~1 a
Pt
Ir
PtÈIr
p(H )/kPa
EUROPT-1
MS3
MS7b
MS4
MS8
MS1
MS2
2
8
16
32
48
56
64
1.0
2.2
4.1
5.1
5.6
2.8
5.6
8.0
9.5
9.5
3
9.5
12
12
3
5
13
16
4.8
9
15.5
21.5
2
5
10
14.5
2
3
8.5
12
12
18
21.5
15.5
13.5
a T \ 603 K, p(hexane) \ 10 kPa, sampling time \ 5 min. b Ref. 16.
3818
Phys. Chem. Chem. Phys., 1999, 1, 3817È3822

View Article Online
Table 2 SpeciÐc activities on Pt and Ir catalysts as a function of the hydrogen pressure
SpeciÐc activity/10~4 mol (mg metal)~1 h~1
p(H )/kPa
Fragments
Isomers
MCP
Hexenes
Benzene
2
(A1) MS3: Pt added during solÈgel preparation of Al O È
2
3
0.34
8
16
32
48
56
0.39
0.90
1.68
2.76
2.95
0.90
1.74
1.44
1.05
0.95
0.31
0.11
0
0
0
0.87
1.74
2.24
2.57
2.57
1.23
2.64
3.14
3.14
(A2) MS7: Pt/Al O (RP) from ref. 16 È
2
3
8
16
32
48
64
0.51
2.57
4.56
4.80
5.76
0.39
2.66
3.60
3.84
3.72
1.26
2.66
2.04
1.68
1.20
0.30
0.03
0.12
0
0.54
1.52
1.68
1.68
1.32
0
(B1) MS4: Ir added during solÈgel preparation of Al O È
2
3
8
16
32
48
64
2.34
4.05
11.2
14.24
16.4
0.12
0.20
0.52
0.56
0.54
0.11
0.20
0.26
0.24
0.18
0.11
0.10
0
0
0
0.33
0.45
1.04
1.12
0.90
(B2) MS8: Ir-Al O (RP) È
2
3
8
16
32
48
64
3.55
7.29
0.12
0.27
0.47
0.32
0.32
0.12
0.18
0.31
0.11
0.11
0.14
0.09
0
0
0
0.86
1.17
1.40
1.08
0.65
13.33
20.00
20.43
only, this catalyst also had a superior isomerization and
(Tables 1È3). As reported previously,1,16,21 the selectivities of
the four ““skeletalÏÏ products were in a similar range (15È30%
each) for Pt catalysts. The values for isomers increased and
(related) C -cyclization activity.
5
those of MCP decreased with increasing p(H ), the selectivities
of benzene and hydrogenolysis being less sensitive to changes
Selectivity
2
SarkanyÏs conclusions concerning Pt and Ir were mainly
based on selectivities.24,27 However, selectivities normalised to
100% may not reÑect properly the real di†erences between
di†erent metals exhibiting very large activity di†erences
in H pressure whereas alkenes appeared at the lowest p(H )
2
2
only. This is essentially similar to the behaviour of Ir catalysts,
apart from the much higher fragment selectivity. Fig. 1 shows
a selected example of Pt and Ir on Al O (RP) together with
2
3
Table 3 Intrinsic activities on PtÈIr catalysts as a function of the hydrogen pressure
Intrinsic activity, TOF/h~1
p(H )/kPa
Total
Fragments
Isomers
MCP
Hexenes
Benzene
2
(A) T \ 603 KÈ
(A1) MS1: PtÈIr co-impregnated on Al O È
2
3
8
16
32
48
64
80
200
430
620
680
24
100
250
440
544
8
24
64
61
48
19
32
34
19
14
10
2
0
0
0
19
42
82
100
75
(A2) MS2: PtÈIr on Al O from bimetallic precursorÈ
2
3
8
16
32
48
64
170
260
640
870
1000
37
96
365
565
880
13
26
51
52
10
7
16
32
35
10
68
52
19
9
44
70
173
209
100
0
(B) T \ 663 KÈ
(B1) MS1: PtÈIr co-impregnated on Al O È
2
3
8
16
32
48
64
190
300
540
640
960
41
105
313
410
884
17
30
43
38
9
8
21
27
19
9
74
60
11
7
50
84
146
166
58
0
(B2) MS2: PtÈIr on Al O from bimetallic precursorÈ
2
3
8
16
32
48
64
270
410
760
1180
1360
48
119
365
684
870
16
41
84
118
108
11
37
76
71
55
125
94
38
24
14
70
119
197
283
313
Phys. Chem. Chem. Phys., 1999, 1, 3817È3822
3819

View Article Online
2.7 ] 10~3. Some of the metal was partially encapsulated into
the support after this preparation, as opposed to the increased
Ir/Al ratio (5.5 ] 10~3) in
a conventionally prepared
Ir/Al O , indicating surface enrichment of the metal. This dif-
2
3
ference manifested itself Ðrst in hydrogenolysis reactions
(Table 2), requiring larger ensembles. The overlapping of the
Pt 4f and Al 2p peaks did not permit such a comparison with
Pt/Al O .
2
3
As opposed to the negative hydrogen exponent reported
earlier,4,19 an increase in hydrogen pressure accelerated the
reaction rates markedly. This can be easily explained by con-
sidering that they exhibit, as a rule, maxima as a function of
p(H ).20,21,23 These maxima correspond to ““optimumÏÏ
2
hydrogen coverages as described in terms of competition of
reactive intermediates and hydrogen for free active sites.28
They are shifted towards larger p(H ) values at higher
2
temperatures20,21,29,30 since higher and higher pressures are
necessary to maintain the optimum coverage. Hence, the con-
tradiction is only apparent: at the lower temperature (around
470 K) Sinfelt and Yates4,19 must have come across the
descending branch of the bell-shaped curve. Sarkany also
reported a maximum in the overall rates of C ÈC alkane
5
6
transformations on Ir/Al O 24 and Ir black31 in the range
2
3
473È543 K. Our higher temperatures corresponded to the
ascending (positive hydrogen exponent) branch on all cata-
lysts, approaching the maximum rate in some cases (Table 1).
Essentially two interpretations have been proposed for
hydrogen pressure e†ects: one claimed that surface H controls
the abundance of various surface intermediates of di†erent
degrees of dehydrogenation.1,20,29,31 This would leave the
ensemble size essentially intact. Sarkany24,27 argued on the
basis of coking experiments with Pt and Ir catalysts that the
ensemble size would change with hydrogen pressure. Hydro-
genolysis would accordingly be favoured on larger ensembles
at high p(H ) whereas low p(H ) would favour the accumula-
Fig. 1 Selectivity of products formed from hexane on Pt/Al O and
2 3
function of hydrogen pressure.
p(Hexane) \ 1.33 kPa, T \ 603 K, sampling time \ 5 min. Note the
expansion of the lower section of the ordinate in the middle panel and
the di†erent scaling in the upper panel.
Ir/Al O
catalysts as
a
2
3
the bimetallic PtÈIr co-impregnated on the same alumina. The
selectivity di†erences of the two PtÈIr catalysts were more
marked at 663 K as shown in Fig. 2.
2
2
tion of adsorbed hydrocarbons and on these, smaller, ensem-
bles non-degradative reactions would prevail.
We think that the two explanations are complementary
rather than contradictory. The higher hydrogenolysis activity
of ““cleanÏÏ Pt surfaces has been demonstrated without
doubt.12,24,25,27 With decreasing hydrogen pressure the pre-
vailing reaction on EUROPT-1 was shifted gradually from
Discussion
The two preparation methods for the Pt or Ir monometallic
catalysts gave rise to di†erent speciÐc activities (per unit mass
of metal), the highest activity corresponding to both conven-
tionally impregnated samples (Tables 1 and 2). Unpublished
XPS results from Poitiers gave a likely explanation for this:
the surface Ir/Al atomic ratio of the catalyst prepared by the
solÈgel process was 1.5 ] 10~3 as opposed to the bulk ratio of
isomerization to MCP formation and, at the lowest p(H ), to
2
dehydrogenation, producing hexenes.23,32 This was attributed
to the gradual accumulation of surface ““hydrocarbonaceous
depositsÏÏ.1,30,33 They suppressed the most active hydro-
genolysis sites. If their amount was moderate and their chemi-
cal state not too dehydrogenated, they were found to be
necessary for steady-state activity in non-degradative alkane
reactions. This state could be maintained by applying a suit-
able hydrogen excess. At the same time, the formation and
rupture of a C ring should involve the same active ensemble
5
but the intermediate may be dehydrogenated to a di†erent
extent. Increasing the hydrogen pressure increased the isomer/
fragment ratio over Pt, Pd and even Rh.20 Hence the very
high hydrogenolysis activity may be a speciÐc property of Ir.
The classical work of Sinfelt4 correlated hydrogenolysis
activity with the electronic properties of various metals. The
negative hydrogen exponent was found to be higher for Pt
than for Ir, hence, the assumed surface intermediate of ethane
hydrogenolysis would accordingly be (C H ) as opposed to
2
0 ads
(C H ) on Ir.4 The number of H atoms in that intermediate
2
2 ads
on various metals was considered to be directly correlated
with the ratio of dehydrogenation to hydrogenolysis activities.
Hydrogenolysis activity was assumed to involve single C
atoms on the metal.10 The strength of the IrÈC bond is higher
than that of the PtÈC bond (621 vs. 608 kJ mol~1),10 although
the di†erence is not large. The high abundance of
““irreversiblyÏÏ adsorbed hydrogen on Ir/Al O 34 would then
Fig. 2 Selectivity of products formed from hexane on two
PtÈIr/Al O
catalysts as
a
function of hydrogen pressure.
2
3
2
3
p(Hexane) \ 1.33 kPa, T \ 663 K, sampling time \ 5 min.
hydrogenate these entities to methane. The CÈC bonds would
3820
Phys. Chem. Chem. Phys., 1999, 1, 3817È3822

View Article Online
Table 4 SpeciÐc activities on PtÈIr/Al O catalysts as a function of the hydrogen pressure
2
3
SpeciÐc activity/10~4 mol (mg metal)~1 h~1
p(H )/pKa
Fragments
Isomers
MCP
Hexanes
Benzene
2
(A) MS1 PtÈIr/Al O (RP) co-impregnatedÈ
2
3
8
16
32
48
64
0.6
2.5
5.8
10.5
12.2
0.2
0.6
1.5
1.4
1.1
0.46
0.8
0.8
0.5
0.3
0.26
0.05
0
0
0
0.46
1.1
1.9
2.2
1.6
(B) MS2 PtÈIr/Al O from bimetallic precursorÈ
2
3
8
16
32
48
64
0.44
0.9
4.2
6.7
8.65
0.22
0.6
1.9
2.5
2.4
0.52
1.0
1.3
1.2
1.1
0.46
0.06
0.0
0.0
0.0
0.36
0.5
1.2
1.6
1.35
(C) Calculated average values from results on Pt/Al O and Ir/Al O (MS7 and MS8, T able 1) È
2
3
2 3
8
16
32
48
64
2.0
5.0
8.9
12.4
13.1
0.26
0.7
1.4
1.2
0.9
0.65
0.22
0.06
0.06
0
0.70
1.35
1.55
1.4
1.47
2.05
2.1
2.0
0
1.0
not break but give C2C bonds on dehydrogenating catalysts
that, in turn, could give coke. We attribute the higher rate of
carbon accumulation on Ir compared with Pt10 to the
““polymericÏÏ route of carbonisation being more important
bimetallic complex. No clear explanation was o†ered that
time, but we think now that the highly oxidised Ir in the
former case must be related to the oxidation of monometallic
Ir particles. This oxidation is considerably hampered in mixed
ensembles of Ir with Pt, so Pt can ““protectÏÏ Ir from being
deeply oxidised. At the same time, the enrichment of Cl on the
metalÈsupport borderline corresponds to previous experi-
ence.37
As for the non-degradative reactions, MS2 from the
bimetallic precursor possessed a higher activity for forming
isomers and methylcyclopentane, as opposed to the higher
aromatization activity of the co-impregnated sample. The dif-
ferent structureÈsensitivity of alkane hydrogenolysis and split-
than coking via the ““C routeÏÏ.35
1
The activity results in Table 3 show that the mixed PtÈIr
catalysts preserved most of the ability of Pt to produce non-
degradative products and the high hydrogenolysis propensity
of Ir was preserved. This indicates a considerable amount of
exposed Ir atoms on the surface; Rasser et al.10 reported
65È70% Pt on the surface with a nominal composition of
Pt : Ir \ 1 : 1. Table 4 compares the speciÐc activity values of
both PtÈIr samples with the average activities calculated from
monometallic Pt and Ir on Al O (RP). The hydrogenolysis
ting of a C ring (a possible intermediate for isomerization14)
2
3
5
activity of both catalysts was well below that of the PtÈIr
average, in agreement with previous reports.3,10 What is new
in this work is the hydrogen pressure sensitivity of the hydro-
should be considered here.38 Hydrogenolysis requiring prob-
ably larger ensembles was less hampered on the co-
impregnated sample than the latter non-degradative process
requiring probably two-atom sites.1 Its larger ensembles also
promoted aromatization, especially with less surface hydrogen
present at 663 K (Fig. 2). We fully agree with the opinion of
Ponec and Bond3 that ““the optimal composition and number
of mixed ensembles may be decisiveÏÏ in determining the activ-
ity shown in di†erent reactions.
genolysis activity of the bimetallic catalyst: at low p(H ) the
2
co-impregnated sample corresponds to the value of the
expected Pt enrichment but at 64 kPa it is almost equal to the
PtÈIr average. This may be in accordance with the large
amount of irreversibly chemisorbed H on Ir,34 activating Ir
ensembles more than Pt sites.
The PtÈIr catalyst from the bimetallic precursor was much
less active in hydrogenolysis, the activity being well below the
PtÈIr average. The e†ect of hydrogen enhancing fragmenta-
tion appeared to a lesser extent here. This may indicate a
more intimate PtÈIr dispersion, similar to the suggestion for a
PtÈSn catalyst prepared from the bimetallic precursor.15h17 A
semi-quantitative EDX analysis conÐrmed this suggestion.
The electron microscope picture of MS1 showed (1) very Ðnely
dispersed metal particles containing mostly Pt but with Ir also
present and (2) larger, elongated crystallites of pure Ir. Sample
MS2, in turn, contained small particles (diameter from about
3 to 50 nm) containing 70È90% Pt, together with larger
agglomerates where the composition changed from spot to
spot, the Pt content being between 30 and 50%. Thus, EDX
conÐrmed the presence of separate Pt and Ir particles in the
co-impregnated sample wheres only bimetallic particles (of dif-
ferent composition) were detected in the sample from the
complex precursor. Of course, EDX can indicate nothing
about the surface composition of metal particles. However, an
EXAFS study of the same samples36 showed the average
A.C. thanks the French Ministry of Foreign A†airs for sup-
porting her stay in Budapest during which the experiments
were carried out. Z.P. is grateful to the University of Poitiers
for inviting him as a guest professor, permitting the manu-
script to be completed. The authors thank the Service Micro-
scopie of the CNRS Institut de Recherches sur la Catalyse,
Villeurbanne, and personally thank Dr M. Nicot for carrying
out the EDX analysis and Dr S. Rossignol and M. I. Kane for
the preparation of the solÈgel samples.
References
1
Z. Paal, in Catalytic Naphtha Reforming, ed. G. J. Antos, A. M.
Aitani and J. M. Parera, Marcel Dekker, New York, 1995, p. 19.
J. M. Parera and N. S. Figoli, in Catalytic Naphtha Reforming,
ed. G. J. Antos, A. M. Aitani and J. M. Parera, Marcel Dekker,
New York, 1995, p. 45.
2
3
V. Ponec and G. C. Bond, Catalysis by Metals and Alloys,
Studies in Surface Science Catalysis, vol. 95, Elsevier, Amsterdam,
1995.
4
5
6
J. H. Sinfelt, Adv. Catal., 1973, 23, 91.
number of next neighbours of Ir to be IrCl
0.6
tion of the co-impregnated sample at 673 K as opposed to the
O after oxida-
B. E. Nieuwenhuys and G. A. Somorjai, J. Catal., 1977, 46, 259.
A. L. Bonivardi, F. H. Ribeiro and G. A. Somorjai, J. Catal.,
1996, 160, 269.
5
composition of IrCl of the catalyst prepared from the
O
1.5
2
Phys. Chem. Chem. Phys., 1999, 1, 3817È3822
3821

View Article Online
7
8
9
M. W. Vogelzang, M. J. P. Botman and V. Ponec, Faraday
Discuss. Chem. Soc., 1981, 72, 33.
23 G. C. Bond and Z. Paal, Appl. Catal. A, 1992, 86, 1.
24 A. Sarkany, J. Chem. Soc., Faraday T rans.,1989, 85, 1511.
25 V. Ponec, Adv. Catal., 1983, 32, 149.
G. B. McVicker, P. J. Collins and J. J. Ziemiak, J. Catal., 1982,
74, 156.
26 Z. Paal and P. Tetenyi, J. Catal., 1973, 30, 350.
27 A. Sarkany, J. Chem. Soc., Faraday T rans., 1989, 85, 1523.
28 G. C. Bond and J. C. Slaa, J. Mol. Catal., 1995, 98, 81.
29 Z. Paal and G. A. Somorjai, in Handbook of Heterogeneous
Catalysis, ed. G. Ertl, H. Knozinger and J. Weitkamp, Verlag
Chemie, Weinheim, 1997, vol. 3, p. 1084.
C. G. Walter, B. Coq, F. Figueras and M. Boulet, Appl. Catal. A,
1995, 133, 95.
10 J. C. Rasser, W. H. Beindor† and J. J. F. Scholten, J. Catal., 1979,
59, 211.
11 F. J. Kuijers and V. Ponec, Appl. Surf. Sci., 1978, 27, 235.
12 J. G. Van Senden, E. H. Van Broekhoven, C. T. J. Vreesman and
V. Ponec, J. Catal., 1984, 87, 468.
30 A. Wootsch and Z. Paal, J. Catal., 1999, 185, 192.
31 A. Sarkany, K. Matusek and P. Tetenyi, J. Chem. Soc., Faraday
T rans. 1, 1977, 73, 1699.
13 Z. Paal and P. Tetenyi, Nature (L ondon), 1977, 267, 234.
14 F. G. Gault, Adv. Catal., 1981, 30, 1.
32 Z. Paal, Adv. Catal., 1980, 29, 273.
15 C. Kappenstein, M. Saouabe, M. Guerin, P. Marecot, I. Uszkurat
and Z. Paal, Catal. L ett., 1995, 31, 9.
33 Z. Paal, H. Groeneweg and J. Paal-Lukacs, J. Chem. Soc.,
Faraday T rans., 1990, 86, 3159.
16 Z. Paal, A. Gyory, I. Uszkurat, S. Olivier, M. Guerin and C.
Kappenstein, J. Catal., 1997, 168, 164.
34 G. B. McVicker, R. T. K. Baker, R. L. Garten and E. L. Kugler,
J. Catal., 1980, 65, 207.
17 C. Kappenstein, M. Guerin, K. Lazar, K. Matusek and Z. Paal,
J. Chem. Soc., Faraday T rans., 1998, 94, 2463.
35 A. Sarkany, H. Lieske, T. Szilagyi and L. Toth, in Proceedings of
the 8th International Congress on Catalysis, Berlin, Verlag
Chemie, Weinheim, 1984, vol. 2, p. 613.
18 B. Michelot, A. Ouali, M.-J. Blais, M. Guerin and C. Kappen-
stein, New J. Chem., 1988, 12, 293.
36 A. El Biyyadh, M. Guerin C. Kappenstein, D. Bazin and H.
Dexpert, J. Chim. Phys., 1989, 86, 1751.
37 K. Hayek, R. Kramer and Z. Paal, Appl. Catal. A, 1997, 162, 1.
38 J. Barbier and P. Marecot, Nouv. J. Chim., 1981, 5, 393.
19 J. H. Sinfelt and D. J. C. Yates, J. Catal., 1967, 8, 82.
20 Z. Paal, in Hydrogen E†ects on Catalysis, ed. Z. Paal and P. G.
Menon, Marcel Dekker, New York, 1988, p. 449.
21 Z. Paal, Catal. T oday, 1992, 12, 297.
22 T. Lopez, A. Asomoza, L. Razo and R. Gomez, J. Non-Cryst.
Solids, 1989, 108, 45.
Paper 9/02976A
3822
Phys. Chem. Chem. Phys., 1999, 1, 3817È3822